How to choose your lab's first challenges

Choose challenges that a named owner already loses sleep over, that can show a visible result within one quarter, and that the lab can actually influence. Your first three challenges are not really about innovation output; they are how the organisation decides whether the lab is worth keeping. Pick a small, balanced set, filter it with explicit criteria, and say no to the rest in writing.

Why the first challenges matter more than the rest

A new lab has no track record. Leaders, middle managers and staff judge it by what it works on first. If the lab chooses a fashionable topic nobody owns, it looks like a side project. If it chooses the hardest problem in the organisation, it disappears for a year. If it chooses something trivial, it looks like a suggestion box with furniture.

The first challenges also set the lab's operating model by accident. Whatever you do first becomes "how the lab works": who brings challenges, who decides, how long things take. Choose them deliberately.

Step-by-step: from long list to first three

1. Collect challenges from owners, not from ideas

Do not start with an ideas campaign. Start by interviewing ten to fifteen directors and frontline managers. Ask about problems, not solutions: "What keeps going wrong? What do customers complain about that you cannot fix? Where does work get stuck?" Stakeholder interviews give you challenges that come with an owner attached.

Add two other sources: existing data (complaints, audit findings, service delays, a Pareto analysis of where most of the pain sits) and the strategy, so that at least part of your list links to what leadership has already committed to.

2. Write each challenge as a problem, not a project

"Build a chatbot" is a solution. "First-time applicants abandon the licence form halfway" is a challenge. Rewrite every item as a problem with a who, a what and a consequence. A problem tree helps separate the visible symptom from the cause you might work on. If you cannot write it as a problem, park it.

3. Filter out what the lab should not take

Before scoring, remove items that fail any of these:

• No owner. Nobody with authority will receive the result.
• Already decided. The solution is chosen and the lab would be decorating it.
• Outside your reach. It depends on a law change, a national system or a budget you cannot touch in the next year.
• Pure delivery. The answer is known and it just needs doing; that belongs in a project office, not a lab.

This step usually halves the list. It is also the step where you learn the most about how the organisation sees the lab.

4. Score the shortlist against explicit criteria

Use a weighted decision matrix with criteria agreed before anyone sees the options. For a first round, these work well:

Weight owner pull and speed to evidence highest in the first round. Later rounds can shift weight to value and strategic link.

5. Balance the set, do not just take the top three

The top three scores are often three versions of the same kind of problem. Plot the shortlist on an impact-effort matrix or a simple portfolio map, and aim for:

• one quick, visible win that shows the lab can deliver;
• one meaty service problem that shows depth;
• one challenge from a part of the organisation that is sceptical of the lab.

The third is the one leaders forget, and it is often the one that earns the lab its allies.

6. Frame each one as a question the team can work on

Turn each chosen challenge into a working question. A good how might we question is narrow enough to act on and open enough to allow more than one answer. "How might we help first-time applicants finish the licence form in one sitting?" is workable. "How might we transform licensing?" is not.

7. Agree terms with each owner before you start

Write a one-page agreement per challenge: the question, the owner, the people they will release, the evidence that would count as success, and a review date. This is also the start of the handover; see from lab to operations.

8. Say no in writing, and keep the parked list

Write back to every person whose challenge was not chosen. Say why, and say when the next intake will be. A parked list that is published and revisited builds trust; a silent rejection builds the story that the lab only works for senior favourites.

Illustrative example

Illustrative. A national university in Central Europe opened a service innovation lab inside its student services division. The lab manager interviewed fourteen heads of department and collected forty-one items. Rewriting them as problems removed nine that were really requests for software. The filter removed twelve more: five had no owner, four depended on national regulation and three were already being delivered by IT.

Of the remaining twenty, the scoring favoured three about the admissions office. The lab manager kept one of them, replaced the second with a problem from the library (where staff had been openly doubtful about the lab) and replaced the third with a quick win: international students missing their housing deadline because the reminder email went to an address they no longer checked.

The housing fix showed results within six weeks. The library challenge took longer, but the library director became the lab's most vocal supporter at the next budget meeting.

Do's and don'ts

Do

• Start from owners and problems, not from an ideas campaign.
• Agree the scoring criteria before anyone sees the list.
• Include one challenge from a sceptical department.
• Get a written owner agreement with released time.
• Publish the parked list and the date of the next intake.

Don't

• Take the CEO's pet idea as the first challenge just because it came from the top; put it through the same filter.
• Pick the organisation's biggest unsolved problem as your first project.
• Accept challenges that are really solutions in disguise.
• Run more than three or four first challenges with a small team.
• Let the scores make the decision; they inform it.

Questions people ask

How many challenges should a new innovation lab take on? For a small lab team, three is a sensible first round, with a fourth only if it is a quick win. More than that spreads attention thin and delays the first visible result, which is what the lab most needs.

What makes a good innovation lab challenge? A real problem with a named owner who wants it solved, a result that can be shown within roughly a quarter, and causes that the lab and the owner can actually change. A clear link to strategy helps but is not enough on its own.

Should an innovation lab accept challenges from anyone? Accept challenges from anyone, but choose them with transparent criteria and publish what was parked and why. Open intake with a clear filter builds trust; open intake with no filter turns the lab into a queue.

Where to go next

Start with the hub piece, what an innovation lab is, and why most fail, and if you are inside a public body read how to set up an innovation lab in a government entity. The guides you will use most in choosing challenges are stakeholder interviews, the problem tree, the weighted decision matrix and how might we.